Computer-Aided Design for Construction
A building is of course an assembly of some in situ construction and many premanufactured standard components. When we consider the vast number of components included in a typical building, the essential aim of design documentation is to identify in each case which component is required, and its precise intended location. We are rarely designing the components themselves. Normally their exact 3-D form does not matter too much and does not have to be shown in much detail. So it is conventional to represent little more than the shape of the component outline in plan provided that this makes the item reasonably recognizable. For example, a wash hand basin only requires its outline shape, and a suggestion of the bowl, in plan view.?If the outline is accurate in size, then the component can be placed on a building plan with reasonable confidence that it will not clash with other elements in the vicinity.
When a CAD system is in use, the components in a building are rarely so tightly packed that a 2-D, or perhaps a box geometry, representation is going to lead to many problems of clashing components. Full 3-D representation might be more necessary in the design of a car where component shapes are more critical. But for buildings, 3-D representation for everything might prove to be wasteful in design effort. The building designer is well able to think in 3-D and eventually the results of his work must be a series of 2-D drawings for the contractor. It does help, however, if the manufacturer's name or other descriptive material can be attached to graphical components within the computer data model, to facilitate the printing of bills of?materials.
It is not possible to generalize too much, however. An exception to the above view occurs when the building is dominated by very highly standardized components, and the component density is high.?For example, hospitals with their high services content can fall into this category. If time and trouble is taken to build up a database containing full 3-D descriptions of the range of standard components to be used, the benefits of being able to design in full 3-D may repay the initial effort. The designer can achieve a higher degree of optimization simply by rearranging components in different ways, and viewing the results from different viewpoints. It is cheaper to deal with impractical layouts and clashing of tightly packed components on a screen, compared with making alterations with the actual components later on site.
A characteristic of buildings is that there are typically tens of thousands of components, although many are similar items repeated many times over at different locations. So when CAD methods are employed, keys to success are:
(1) Simple representation (frequently symbols are adequate)
(2) Very rapid positioning of components?
(3) Accurate positioning of components
(4) Ease of copying components
(5) Ease of moving previously placed components
(6) Use of parameterized shapes, i. e. the same basic components that occur in different sizes, for example, steel sections, pipes, ducts, radiators.
In some CAD systems the act of copying a component each time involves the duplication of its entire graphical data within the database, as well as storing the new position. In use for building design, such databases rapidly become large, cumbersome and use up?much computer power. Systems more suitable for building design would merely store the new component position together with a reference pointer back to the original single set of graphical data. This is much more concise.
The large number of components and the variety of disciplines makes a sophisticated layering or component selection system very necessary.
When a designer works on a drawing board, he may be working either on a general arrangement at small scale, or on details at large scale. Large components tend to be shown on the general arrangement but the joints between them are on detail drawings. What tends to get missed out in this process, only to reappear as headaches on site, are the problems of fit between different items.
However in CAD operation the concept of drawings and scale tends to remain rather in the background until after the design is basically finished, and the drawing sheets have to be plotted. Before that time it is the data model that rules. In 2-D working this may be an entire floor with all construction and components. In box geometry it is the whole building. So element after element can be placed with considerable precision. If there are large precast concrete cladding panels, then these can be placed side by side with the correct space left for jointing and tolerances. This of course is how it is attempted manually with the GA drawing. But now the designer can zoom in as much as he likes with a screen to explore the problems of fit between unlike components placed perhaps at different times by different designers. So the gulf between general arrangement design and detail design is rather less apparent. This is a major benefit.
Some thought needs to be given to how certain detail will ultimately be presented on drawings. At small scale little may be shown of any item or component apart from an outline. On large scale?drawings, more detail may be required. Presentation for each instance can be made easier if the more limited graphics is included in one category or layer, and the greater detail in another. Then the choice for particular drawings can be made simply by the user turning on the relevant categories for each display or plot.